I need to do a rolling sum of a columns of a dataframe, verify if the sum of the rolling (2 to 5 lines) is = 0 and take all the lines from the rolling to flag them as "canceled".
I tried this :
df_sorted = df.sort_values('No Piece FI')
df_sorted['No Piece FI'] = df_sorted['No Piece FI'].astype(str)
df = df_sorted[df_sorted['No Piece FI'].str.startswith('11')]
rolling_sum = df['CP'].rolling(window=5, min_periods=2).sum().shift(-4)
df.loc[(rolling_sum == 0), 'Annulé'] = 'Canceled'
But when I do that the result didn't seems to works. Also I would like to see the lines number like that or at least the group :
| Annulé | Lines | CP | Group (optionnal) |
|---|---|---|---|
| Not canceled | 8 | ||
| Not canceled | 4 | ||
| Canceled | 1 | 10 | 1 |
| Canceled | 2 | -10 | 1 |
| Not canceled | 3 | ||
| Canceled | 1 | 15 | 2 |
| Canceled | 2 | -5 | 2 |
| Canceled | 3 | -10 | 2 |
| Canceled | 1 | 8 | 3 |
| Canceled | 2 | -8 | 3 |
| Not canceled | 19 | ||
| Not canceled | 3 | ||
| Not canceled | 2 | ||
| Canceled | 1 | -8 | 4 |
| Canceled | 2 | 2 | 4 |
| Canceled | 3 | 2 | 4 |
| Canceled | 4 | 1 | 4 |
| Canceled | 5 | 3 | 4 |
| Not canceled | -1024 |
Here is what only the input would looks like :
| CP |
|---|
| 8 |
| 4 |
| 10 |
| -10 |
| 15 |
| -5 |
| -10 |
| 8 |
| -8 |
| 19 |
| 3 |
| 2 |
| -8 |
| 2 |
| 2 |
| 1 |
| 3 |
How can I do that ? Thanks you.
